Privacy Preserving Data Mining (PPDM) enables one or several
parties to conduct the normal data mining operations
without knowing the data (or without knowing some part
of the data). PPDM recently has received more and more attention
because of the increasingly important privacy concerns.

In general, there are two approaches for PPDM problems. One is the
randomization approach, in which, data are randomized before being
disclosed to the data miner.  The second approach is the secure
multi-party computation approach, in which, data are disguised using
cryptographic methods to allow multiple parties to conduct a joint
computation without disclosing one party's private data to the others.

In this talk, I will give an overview of the PPDM problems
and both approaches. I will present some of my past and
current work in this area.

Considering that some students might not have background in
data mining, I will present the work in a way that requires
no background in data mining.